To the Editor We read with interest the article by Tripathy et al 1 reporting the results of the ATTAIN randomized clinical trial, which found no statistically significant difference in survival outcomes between treatment with etirinotecan pegol (EP) and chemotherapy of physician's choice (CPC) in patients with breast cancer with brain metastases. Although this study is impressive, we have a few concerns and suggestions for this study.This study 1 did not compare the size, number, and location of brain metastases between the 2 groups to control confounding factors. This study also did not report the comparison of objective response rate (ORR) between the EP and CPC groups. Table 2 showed that the ORR was extremely low and similar between the 2 groups. However, the high percentage (range, 20.5%-41.1%) of not-evaluable cases may induce bias for response analysis. In light of the results of the BEACON trial, 2 this study 1 may select the ORR as the primary end point and survival outcomes as the secondary end point, as a prior study explored the efficacy of EP in refractory ovarian cancer. 3 The authors 1 stated that there was no difference in median progression-free survival by blinded independent central review; yet Table 2 displayed that the median progression-free survival for central nervous system plus non-central nervous system metastases per blinded independent central review for EP vs CPC was statistically different (P = .02). Considering this study included 7 chemotherapy drugs selected by physicians, it would be significant to compare EP with single-agent chemotherapy one by one.It is too early to tell whether EP is ineffective in the treatment of advanced cancer. We can leverage the power of artificial intelligence (AI) based on historical clinical trial data and real-world data to optimize the clinical trial design and improve the success rate. Some recent studies have used AI based on biomarkers as the inclusion criteria to select "the ideal" patients for clinical trials, 4,5 and AI models also can predict drug response to reduce clinical study sizes and improve clinical trial performance. We can also use AI to create a synthetic control arm to make trials more patient-centric, shorten enrollment timelines, and increase power and confidence.Although this study 1 could not bring positive results for EP, in the future, it is possible that we can apply AI to identify the subset of patients who likely benefit from EP and combine EP with other novel biologic agents for better treatment outcomes. In such a setting, participation in clinical trials is strongly encouraged.